PSA screening in prostate cancer: The controversy continues

September 9, 2020

A new analysis supports return to widespread screening.

A reevaluation of the tradeoffs of prostate cancer screening using the most recent data should rewrite the negative narrative associated with PSA testing, according to an analysis published in the New England Journal of Medicine (NEJM).1

In their paper, the researchers explain that PSA screening became a widespread practice in the early 1990s, and over the ensuing 25 years, there was a significant increase in the number of prostate cancer diagnoses, along with an approximately 50% decrease in mortality rates. However, over the past 10 years, PSA screening has significantly declined, as initial findings from randomized trials cast doubt on its actual value and professional organizations began issuing guidelines recommending against screening.

In 2012, the US Preventive Services Task Force (USPSTF) issued a grade D recommendation against the use of PSA screening in the general US population, regardless of age. The current USPSTF PSA screening policy is slightly changed, with a grade C recommendation for men aged 55 to 69 years, meaning in this population, an individual decision on screening should be made based on a physician-clinician discussion of the potential benefits and risks.

The recommendation notes that some men in this population may have a positive benefit-risk profile with screening, but warns of potential harms, including false-positives, overdiagnosis/overtreatment, and treatment-related adverse events, such as erectile dysfunction and incontinence.

The USPSTF based this recommendation for men aged 55 to 69 on its own review of the literature. The panel concluded that PSA screening in this population can prevent an estimated 1.3 men from dying of prostate cancer every 13 years per 1000 men screened. The review also determined that PSA screening may prevent approximately 3 cases of metastatic prostate cancer per 1000 men screened.

For men aged ≥70 years, the USPSTF recommends against any PSA-based screening (grade D recommendation). The panel also does not recommend screening for men aged <55 years. While the USPSTF acknowledges that family history and African-American race are two of the most significant risk factors for developing prostate cancer, the panel maintains that there is insufficient evidence to issue individualized screening recommendations in any age group based on those factors.

In the NEJM paper, the researchers maintain that there has been a misunderstanding of the data from 2 pivotal screening trials, leading to public uncertainty on the value of PSA screening and supporting the anti-screening recommendations by panels such as the USPSTF. The misunderstanding includes both a “misinterpretation” of data, as well as not focusing enough on the length of follow-up when determining the benefit-risk profile of screening.

The first trial the paper draws attention to is the US-based Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. The PLCO results showed no benefit for screening; however, subsequent analyses severely questioned the validity of the trial because of mass contamination of the control arm. The trial was intended to compare screening with a control arm of no screening; however, it was determined that nearly 90% of the control arm had received screening before or during the trial, making the study, essentially, a screening versus screening trial with no capacity to measure the value of PSA testing.2

The other trial, the European Randomized Study of Screening for Prostate Cancer (ERSPC), did show a benefit in prostate cancer mortality, but the USPSTF did not consider the benefit sufficient enough compared to potential harms to support screening. The NEJM paper maintains, however, that as the length of ERSPC follow-up grows increasingly longer, the benefit of screening is increasingly revealed.

The latest ERSPC data show that with 16 years of follow-up, 570 men aged 55 to 69 years would need to be screened to prevent 1 death from prostate cancer, with an estimated number of 18 excess diagnoses. For perspective, the NEJM paper points out that for breast cancer screening, for which there is widespread support, 1250 women aged 50 to 59 years need to be screened to prevent 1 death from breast cancer at 10 years. The number needed to screen (NNS) is 476 and 769 women for the age ranges of 60 to 69 years and 70 to 74 years, respectively.

The researchers concluded that a follow-up time of 16 years post-randomization is an insufficient window to capture the complete mortality benefit of PSA testing, noting that while screening often starts for men when they are in their 50s, the median age is 80 years for men who die of prostate cancer. Thus, they used a statistical model to project what the ERSPC results would likely show at 25 years’ follow-up. The model projected that at 25 years’ follow-up, the NNS to prevent 1 prostate cancer death drops to 385, with 11 excess cases.

The researchers also point out that the current USPSTF recommendation does not factor in the evolution of management strategies in prostate cancer, such as the now widely used approach of active surveillance for low-risk disease, which has the potential to mitigate potential harms associated with overdetection.

Overall, while the reduction in prostate screening has led to a decline in prostate cancer diagnoses, emerging evidence is suggesting that the rates of metastatic disease at diagnosis are increasing, whereas they had been on the decline until a decade ago.

Given the increased benefit shown by their model and other emerging factors in the PSA-screening picture, the researchers strongly believe the USPSTF should change its position on PSA screening.

“My thought, based on our work, is that it should be recommended for all healthy men to undergo some sort of PSA screening,” study author Jonathan Shoag, MD, assistant professor of urology at University Hospitals Cleveland Medical Center, Case Western Reserve University School of Medicine, Cleveland, Ohio, and adjunct assistant professor of urology at Weill Cornell Medicine, said in an interview with Urology Times®.

“In the context of our study, I think not only is there a benefit to PSA screening, but I think we’re already starting to see some real-world data to suggest what happens when we don’t screen as aggressively as we can,” study author Yaw Nyame, MD, MBA, a clinical fellow in the department of Urology at the University of Washington School of Medicine and Seattle Cancer Care Alliance, and a postdoctoral fellow at the Fred Hutchinson Cancer Research Center, Seattle, Washington told Urology Times® in an interview.

In interviews with Urology Times®, Shoag, Nyame, and senior study author Jim C. Hu, MD, MPH, who is the Ronald P. Lynch Professor of Urologic Oncology at Weill Cornell Medicine and director of the LeFrak Center for Robotic Surgery at NewYork-Presbyterian/Weill Cornell Medical Center, New York City, New York, discussed the history and current state of PSA screening, their paper reevaluating the benefit-risk profile of screening, and why the USPSTF screening recommendation should be changed.

What is the prevailing opinion on PSA screening based on the latest understanding of the PLCO and ERSPC trials? What were some of the issues with those trials?

Shoag: The story is that in the early 90s, PSA screening became pretty widely adopted in the United States. And we realized that we were finding a lot of prostate cancer—much more than before—but also that mortality declined from prostate cancer; there was a 50% drop in mortality

over the subsequent 30 years. However, since we were finding a lot more prostate cancers, people started to ask, “Is PSA screening really beneficial?” And we had 2 large randomized studies of PSA screening—one in the United States, which was the PLCO trial, and one in Europe, which was the ERSPC trial.

The PLCO trial showed no benefit for PSA screening, while the ERSPC trial showed a 20% reduction in mortality from prostate cancer with PSA screening; however, the initial ERSPC follow-up was under a decade, so not many men died of prostate cancer overall, so the absolute benefit was relatively small.

And so that was sort of where we stood, and I think as a result of the idea that we weren’t saving many lives, even in the positive trial, and we had this negative trial, the opinion became, “Well, we shouldn’t do PSA screening anymore.”

In 2016, we looked at the PLCO data, and it turns out that in the PLCO trial, almost all the guys in the control arm had PSA screening, because everyone was getting PSA screening at the time in the United States. So, the PLCO trial really couldn’t inform anything. You can’t assess screening versus no screening in a trial where the control arm had screening.

And it turns out that with the ERSPC trial, actually when you look as time gets longer and longer from the start of the trial, the relative mortality reduction seems to be about the same, but the number of men saved per man screened increases because as you have more follow-up, more men die of prostate cancer, which seems obvious, but unfortunately, it hasn’t been obvious to the people making the screening recommendations.

So, as a consequence of all of this, PSA screening has really fallen out of favor—we’re diagnosing a lot less prostate cancer and treating a lot fewer patients with prostate cancer, and we’re starting to see the epidemiologic consequences of that where we’re seeing more metastatic disease mortality, which had been decreasing since the advent of PSA screening and is now actually increasing, or at the least, is relatively stable. But definitely the decrease seems to have stopped and it’s concerning. And so that’s where our study came in—we said, these things that I’m saying, which seem like obvious issues, but apparently are not obvious to everyone else. So that’s why we wrote this paper.

Nyame: I think when you look at PSA screening, the introduction of PSA in the mid-80s was followed by very rapid adoption and widespread use in the community. So I think when it came time to conduct the US randomized trial (PLCO), you had a large population in the country that

was getting tested. And so when you think about the most ideal scenario for designing a trial that’s looking at the efficacy of screening, you really want to be comparing screening to no screening. And right off the bat, you’re already comparing a group of men with this intervention, where everyone essentially, or a large percentage of the population, had already been screened.

And then, muddying the water more, is the fact that during the trial, those that were supposed to be in the control arm for PLCO actually received PSA testing. So, I think I’ve heard people describe that trial as, “screening versus maybe a little bit less screening.” But the truth is, it’s really a screening versus screening trial. But it formed the basis of guideline recommendations for what we should do about PSA, which have really stuck for almost a decade now. In 2018, the recommendation changed, but the change wasn’t a strong one, it’s still a pretty soft recommendation for a discussion between providers and patients.

What was the approach/methodology for your analysis?

Shoag: We sort of said, “The PLCO trial can’t really inform screening versus no screening.” And our collaborators from the University of Washington had done an analysis before using a little bit more involved statistics. The crux of the analysis was, “If you used a control group with PLCO that did not have screening, what would the outcomes look like?” And that actually found that under that circumstance, the PLCO trial may have shown a benefit for screening.

What our analysis did was that we took the data from the ERSPC and said, “What happens if we project this out to 25 years? What is the benefit of screening in terms of number needed to diagnose and number needed to screen?” And it turns out that those numbers are actually pretty compelling; the numbers needed to diagnose to prevent 1 death—those numbers are pretty compelling and we used pretty conservative estimates. We tried very hard to not overestimate the effect of screening. And even with those very conservative numbers, the benefits of screening still seem, to me—this is a judgment call whether the benefits are greater than the harms—but to me, the benefits were even greater than the harms.

Nyame: I think it’s important to recognize a couple of things. The type of technique we used in our study was previously used to assess whether or not screening was beneficial in PLCO. And that was done in an older study [which showed that] when you account for all of that contamination—people who got screened who were supposed to be on the non-screening arm—the benefit of PSA was about the same in PLCO as in the ERSPC.

For our study, what we were really interested in was this infographic that had been floating around that suggested that in order to prevent 1 prostate cancer death, you had to detect and treat 100 men, and that number was generated from early PLCO data. So the assumption there was that if a man gets screened, that the benefit from screening occurred early and didn’t continue to accrue. And so one of the things we really aimed to do with this model was to not only update those numbers, which are widely cited and used in pictorial depictions of the benefits/harms of prostate cancer screening, but we also wanted to demonstrate the fact that the longer a man lives who’s been screened, the more benefit he gains from PSA testing.

So, when we were able to project numbers out to 16 years, our model was able to match the results from the European trial, which is what we calibrated the model to do, so that was kind of a safety check. But when we took the model out to 25 years, what we found was that instead of 100 men needing to be diagnosed and treated [as suggested by the analysis of the early PLCO data], only 11 men needed to be diagnosed and treated to prevent 1 prostate cancer death.

Hu: We projected the long-term benefits of PSA screening, using data from the only randomized trial without significant contamination, and projected that at 25 years after initiation of PSA screening, the benefit was more favorable than that for mammography for breast cancer, which has a USPSTF grade B recommendation for women.

Where do we go from here?

Shoag: First of all, the message to patients needs to change. For a long time, including now, the screening aids and information available to patients and primary care doctors is misleading. The current screening recommendations from the USPSTF and other panels are [based on] about 13 years of follow-up [from the pivotal screening trials], which is too short. If a man begins screening in his 50s, and he’s not going to die of prostate cancer until his 80s, 13 years is just not enough time. There’s no pretense that that’s enough time to estimate the benefits. So I think we need to look over a patient’s lifetime of what the benefit of screening is for an individual.

The framing of the harms and benefits of PSA screening has been misleading too, because lots of men get prostate cancer, even in the absence of screening, and that point hasn’t been made clearly enough. I think we tried very hard to make that point. Lots of men are diagnosed with prostate cancer, even if we don’t do PSA testing. And I think it’s important to sort of recognize that PSA testing is detecting a lot of these cancers simply earlier at a time when they can be treated.

And I think those 2 messages are sort of the points of our manuscript, and when you look at things in that lens, the benefits are significantly more compelling as compared to the harms, and then the judgment of whether we should treat everyone or not is beyond the scope of this current work. I can give you my thoughts, but those are just my opinions. But whoever is making these decisions, needs to look at things over the longer term and that’s important.

Nyame: I think what’s really challenging for me as a prostate cancer researcher, and I think the rest of the team would agree, is that the narrative around PSA screening is really entrenched in a whole generation of providers. We have the gatekeepers, who are primary care physicians, who are being taught that this is a test that not only doesn’t decrease mortality, but only leads to the over-detection of cancers that don’t affect people’s lives. But we know from the data and from our analysis that’s not true. And so I think what we’re hoping for is that, by providing this high-quality analysis and outcomes that are driven by real-world data, that these projections can help people understand what the true benefit of PSA screening is.

And the other thing we spent a lot of time discussing in the article was overdetection and overtreatment are things that are changing rapidly in the prostate cancer landscape. So I think we can’t use the same lens that we have had for prostate cancer screening and treatment from 2 decades ago for our contemporary practice. We have a lot more tools for trying to be more discriminatory in how we diagnose clinically significant cancers. And certainly the mainstay of management for men who have nonaggressive cancer is to observe their disease.

What is the current USPSTF recommendation and what is your opinion of it?

Shoag: It’s grade C, meaning a provider should have a discussion with their patients about the risks and benefits. In my opinion—and this is an opinion, this isn’t data based—I think that’s unfair, because PSA screening is complicated. And I think the idea that patients and providers are going to be able to make this decision in a short visit in an informed way is naïve. I think an expert opinion needs to decide whether or not we think screening is good. And at that point, recommend it or do not recommend it. My thought, based on our work, is that it should be recommended for all healthy men to undergo some sort of PSA screening, and obviously men could choose not to do that, but [in general] we recommend lots of screening tests to people and, in my opinion, PSA screening should be recommended.

Nyame: What I think is interesting is when you look at the USPSTF recommendation for mammography and screening in breast cancer, it is a grade B recommendation, but [prostate cancer screening has a lower recommendation] despite the fact that PSA screening has very similar numbers [regarding the number needed to screen], albeit over a different time frame—the benefit for PSA screening is over a longer period of time being observed than breast cancer.

Prostate cancer had a grade D recommendation until 2018, and currently has a grade C recommendation. I think what’s perhaps most interesting in the backdrop of all of this is a study recently published in the Journal of the National Cancer Institute showing that the number of metastatic prostate cancer cases has been increasing from the time of the USPSTF recommendation. So in the context of our study, I think not only is there a benefit to PSA screening, but I think we’re already starting to see some real-world data to suggest what happens when we don’t screen as aggressively as we can.

Hu: The USPSTF should re-evaluate screening recommendations from [its current position of] individualized choice, grade C.

The harms of screening have been mitigated. Men do not go straight to biopsy with an elevated PSA. They may now undergo other biomarkers, such as 4K or MRI. Additionally, if a man is diagnosed with prostate cancer, he may opt for active surveillance, which had strong uptake over the past 5 years. I believe the USPSTF recommendation to be outdated due to these recent practice pattern changes. Additionally, the majority of men are not able to make an “individualized choice” and primary care doctors are overwhelmed and may not be aware of the nuances of PSA screening.


1. Shoag JE, Nyame YA, Gulati R, Etzioni R, Hu JC. Reconsidering the trade-offs of prostate cancer screening. N Engl J Med. 2020;382(25):2465-2468. doi:10.1056/NEJMsb2000250

2. Shoag JE, Mittal S, Hu JC. Reevaluating PSA testing rates in the PLCO trial. N Engl J Med. 2016;374(18):1795-1796. doi:10.1056/NEJMc1515131